B. Buecker, Buecker & Associates, Lawrence, Kansas (U.S.)
The refinery, petrochemical, chemical and other manufacturing industries have seen and continue to see many retirements of senior technical employees, some of whom have been with their plants since startup. New personnel are often confronted with challenges that require knowledgeable and quick action, which they may have difficulty providing without thorough training. A very important—and sometimes overlooked—unit operation within these facilities is makeup water treatment for steam generators and other processes. Many cases have been documented in which makeup system upsets have led to boiler tube failures that partially or completely shut down unit processes. Note: Cooling water systems are also often neglected to the detriment of plant operations. Parts 1 and 2 (September 2024) of this article serve as a guide for those employees who may be on a steep learning curve when it comes to makeup system understanding.
General notes. Several fundamental guidelines are appropriate to ensure proper makeup system selection, operation and response to upset conditions. These include:
Raw water characteristics. For decades, surface water and—in some locations—groundwater, have served as the makeup supply for industrial plants. The chemistry of these supplies is greatly influenced by weather, topography, geology and, in some cases, proximity to the ocean (FIG. 1).
The primary elements in the earth’s crust and their percentage mass are:
Silicon is almost exclusively bound with oxygen in nature, sometimes as silica (SiO2 ) but usually with other metals (e.g., aluminum, sodium) as silicates. Also common are significant limestone deposits composed of calcium carbonate (CaCO3 ) with some magnesium carbonate (MgCO3 ). Many of these compounds are only slightly soluble, but they will still dissolve sufficiently to influence makeup treatment design. The physical and chemical characteristics of any surface water supply (TABLE 1) will fluctuate considerably over time, with both short-term and seasonal changes.
Samples with high hardness and alkalinity values are likely located in regions with substantial limestone deposits near the surface.1
Large fluctuations in suspended solids concentrations are very common in surface supplies, particularly rivers, following heavy precipitation. The typical requirements to prepare surface water for downstream use include screening to remove large debris, biocide feed to control microbiological fouling, and clarification/filtration (or sometimes filtration alone) to remove suspended particulates. Activated carbon filtration may also be necessary for supplies that have high concentrations of dissolved organics.
Far too often, engineers rely on a one-time raw water grab sample for treatment system design. This may lead to poor performance and sometimes even failures that require substantial or full replacement of treatment systems. For operating systems, at a minimum, monthly samples are necessary to evaluate changes in natural water chemistry. Parameters should include those in TABLE 1 as well as total suspended solids (TSS), turbidity and total organic carbon (TOC).
In contrast to surface water, many (but not all) groundwater supplies are essentially free of particulates, organics and microbes, as these are removed when the water percolates through soil and rock formations. However, groundwaters often have elevated levels of hardness, bicarbonate alkalinity, silica, iron and even some hydrogen sulfide (H2S), depending on location. Quarterly sampling of groundwater is usually sufficient, and the chemistry of aquifers deeper than 300 ft generally remains consistent within ±10%. Deep wells may have quite high concentrations of hardness, alkalinity and silica, as well as minor but potentially problematic concentrations of barium and strontium.
Fresh water supply issues are forcing personnel at an increasing number of plants to select an alternative makeup source. The most common is effluent from a municipal wastewater treatment plant, also known as a publicly owned treatment works (POTW). These supplies typically contain elevated concentrations of organic compounds, nitrogen species including ammonia and nitrite/nitrate, phosphorus (as phosphate) and suspended solids.
PRETREATMENT
Beyond mechanical screens and settling basins to remove large debris, a core pretreatment process at many facilities is clarification/filtration. A primary goal is to reduce suspended solids that would otherwise rapidly foul downstream equipment, including high-purity makeup systems such as reverse osmosis (RO) and ion exchange units. For waters with high hardness, alkalinity and (at times) silica, lime or lime/soda ash clarification may be necessary. The next sections review these processes.
Suspended solids clarification. Suspended solids, besides being light in weight, typically have a negative charge and are therefore mutually repellant. The initial clarification step is the addition of coagulant (positively charged compounds) to reduce the negative charge on the solids. Coagulation is followed by flocculant treatment to coalesce the particles. Flocculants are polymers that “bridge” coagulated solids into larger particles. The processes are typically performed in a clarifier, with a once-common design shown in FIG. 2.
The coagulant is added to the influent of the “draft tube/mixing zone” for rapid mixing of the chemical to enhance coagulant-solids contact. The discharge from this zone flows to the flocculation zone with a lower velocity to allow the formation of larger flocs. The flocs then enter the “clarification zone” to form a “blanket” that continually collects newly-formed flocs. Periodically, some of the sludge must be blown down to maintain the correct bed level.
Inorganic coagulants are common and include aluminum sulfate [alum, Al2(SO4)3•18H2O], aluminum chlorohydrate (ACH, Al2ClH7O6 ), sodium aluminate (Na2Al2O4 ), ferric chloride (FeCl3 ) and ferrous sulfate [Fe2(SO4)3•9H2O]. These compounds produce highly charged ions, either Al+++ or Fe+++, which then generate gel-like precipitates of Al(OH)3 or Fe(OH)3 that agglomerate solids. Coagulation/flocculation may remove some organic compounds, including those that add color to water. A case-by-case evaluation of coagulants is necessary to identify the most effective per the inlet water chemistry.
Inorganic coagulants offer advantages and disadvantages, as shown in TABLE 2.
Organic, polyelectrolytic coagulants such as poly-DADMAC (diallyldimethylammonium chloride) and polyamine are also available. These compounds can:
Organic coagulants do not add solids to the system or alter the pH and alkalinity.
All coagulants, whether inorganic or organic, are positively charged. Overfeed can seriously affect downstream equipment, particularly RO membranes. These membranes typically have a negative charge that rapidly attracts cationic coagulants, resulting in irreversible fouling.
Flocculants are long-chain polymers that, as noted, act as bridging agents for the coagulated particles. Flocculants may have cationic, anionic or nonionic charge density, with selection based on coagulated water conditions. Many organic flocculants are copolymers of acrylamide, and may include nonionic, cationic or anionic monomers.
In contrast to coagulant feed, flocculants need gentle mixing to influence the process but prevent the flocculated particles from breaking apart. So, it is crucial to feed flocculants to the slow mixing zone in the clarifier.
Polymer manufacturers offer flocculants as powders, emulsions or brine liquids. Floc powders contain approximately 90% active solids. Emulsions and brine liquids have 20%–50% active solids, while pre-diluted floc powders are limited to < 2% active solids. The pre-diluted flocculant option is usually only economical for low-flow applications. In emulsion formulations, the polymers are often curled rather tightly and need time to unwind to be fully effective. One technology to accomplish this is a small, external dilution mixer that uncoils the polymer, which is then fed to the clarifier.
Feed control of coagulant and flocculant is crucial for proper clarifier performance. Under- or overfeeding influences the overall turbidity and solids removal.
FIG. 3 illustrates a clarifier with poor chemistry control that had floc carryover to the effluent followed by the same unit where the coagulant/flocculant chemistry had been optimized.
Changes in raw water flowrate, temperature, chemistry and other factors influence clarifier efficiency. Regular jar testing (FIG. 4) is needed to maintain optimal performance.
A very important clarifier term is the rise rate. This is the flowrate of the clarifier effluent [gallons per minute (gpm)] divided by the surface area of the clarifier (ft2). For clarifiers of the type shown in FIGS. 2 and 3, the rise rate might be at or near 1 gpm/ft2. The footprint for such low rise-rate clarifiers can be quite large. Modern clarifiers offer much larger rise rates with corresponding smaller footprints. One such systema includes a ballast feed and a recycle loop. The unit has both rapid and slow mix zones per normal coagulation and flocculation methods, but a key feature is the injection and recirculation of micro-sand. The sand particles gather the flocs, but being significantly heavier, settle out rapidly in the main clarifier with its lamella plates. Rise rates of 25 gpm/ft2 or greater are possible with ballasted systems. Sludge bleed treatment with a hydrocyclone is an integral part of the process. The centripetal motion imparted by the hydrocyclone causes the heavier micro-sand particles to exit through the underflow and back to the process, while the liberated floc particles exit the overflow for final disposal. This process recovers most of the micro-sand, thus greatly reducing fresh additions.
Other similar systems have evolved, but with important differences. A prime example is another processb with a ballast material of the iron oxide, magnetite (Fe3O4 ). This process has emerged as an effective treatment for some industrial wastewaters containing heavy metals. Some of the metals co-precipitate well with magnetite, and therefore are directly removed from the solution.
Lime softening clarification. Suspended solids clarification alone is often sufficient for surface waters. However, for supplies with high hardness, alkalinity and silica, more complex treatment may be needed, as these compounds can generate scale in high-purity makeup treatment equipment, cooling water systems and other water networks. Cold lime softening (CLS) clarification is an answer for these raw waters.
Coagulant/flocculant feed is still typical for suspended solids removal in CLS clarifiers, but the dominant feature is hydrated lime [Ca(OH)2 ] treatment to remove scale-forming ions and compounds. As a precursor to examination of the fundamental CLS process, FIG. 5 shows how the primary ions in water associate.
The hardness that associates with bicarbonate alkalinity is known as “carbonate” or “temporary” hardness. Hardness associated with chloride or sulfate is known as “permanent” hardness. Sodium and potassium do not form scale, and are not considered in this chemistry.
The addition of lime removes dissolved carbon dioxide (CO2), temporary calcium hardness and magnesium. The main reactions are shown in Eqs. 1–5:
CO2 + Ca(OH)2 → CaCO3↓ + H2O (1)
Ca(HCO3)2 + Ca(OH)2 → 2CaCO3↓ + 2H2O (2)
Mg(HCO3)2 + 2Ca(OH)2 → Mg(OH)2↓ + 2CaCO3↓ + 2H2O (3)
MgCl2 + Ca(OH)2 → Mg(OH)2↓ + CaCl2 (4)
MgSO4 + Ca(OH)2 → Mg(OH)2↓ + CaSO4 (5)
The three fundamental steps of this chemistry are:
Calcium non-carbonate hardness that is in the raw water and is created from reactions 4 and 5 above can be reduced with soda ash (Eqs. 6 and 7):
CaCl2 + Na2CO3 → CaCO3↓ + 2NaCl (6)
CaSO4 + Na2CO3 → CaCO3↓ + Na2SO4 (7)
The need for soda ash feed in addition to lime depends on the raw water composition and the treated water hardness requirements. Caustic (NaOH) and magnesium oxide (MgO) may also be employed to assist with the softening reactions.
A common method for calculating if lime and soda ash dosages are correct uses “P” and “M” alkalinity titration calculations. FIG. 6 outlines the alkalinity relationship in natural waters.
Alkalinity can be easily calculated by acid titrations with two color indicators. The first is phenolphthalein. If a red or pink color develops when the indicator is added, carbonate and possibly hydroxide alkalinity are present. With acid titration, the color disappears at pH 8.3. This is the “P” (phenolphthalein) alkalinity endpoint. The next step is to add the bromcresol green/methyl orange indicator, followed by additional acid titration until the color changes from green to pinkish-red at pH 4.3. This is the “M” (total) alkalinity endpoint.
The theoretical ideal 2P – M value for cold lime softening is 15 ppm. A practical range is 5 ppm–15 ppm. For waters that require the removal of calcium non-carbonate hardness, a slight excess of carbonate is necessary to ensure that the reactions shown in Eqs. 6 and 7 reach maximum completion. In this case, per “P” and “M” alkalinity readings, Eq. 8 calculates the optimal carbonate residual:
2(M – P) = CO3–2 (mg/L) (8)
A common 2(M – P) guideline is 20 mg/L–40 mg/L.
Besides “P” and “M” alkalinity testing, pH, calcium and magnesium measurements are also helpful for monitoring CLS performance. These analyses are straightforward.
CLS effluent is typically unstable and can undergo additional precipitation in storage tanks or filtration equipment. Lowering the clarifier effluent pH to around 9 with sulfuric acid feed or CO2 injection improves effluent stability and reduces the scaling potential.
Silica will combine with magnesium hydroxide in the precipitates formed by reactions (Eqs.) 3, 4 and 5. One ppm of silica combines with 15 ppm of magnesium hydroxide precipitate. This reaction can be very beneficial with high silica waters, where, for example, the silica would “cycle up” to unacceptable concentrations in cooling towers. In some cases, the supplemental addition of magnesium oxide is necessary to achieve the required silica removal.
CLS greatly increases the sludge volume of clarifiers, which in turn complicates sludge disposal requirements. A common feature of CLS systems is a filter press to produce a sludge “cake” that can more readily be disposed.
FILTRATION
While a well-operated clarifier removes most of the suspended solids from the inlet, some solids will still carry over. Even water clarified to 3 NTUs (nephelometric turbidity units) still contains enough suspended solids to foul downstream RO and ion exchange units. Therefore, filtration is an indispensable part of the pretreatment system. This section explores the two basic types of filtration: gravity and pressure.
Gravity filtration. Gravity filtration is often the best choice for high-turbidity waters or waters with high color from organic compounds. Flowrates are slower than for pressure filters, but can potentially reduce effluent turbidity to less than 0.1 NTU.
Filters are often of the multi-media type (FIG. 7) with anthracite, sand and garnet as the three primary layers.
Two varieties of gravity filters, based on backwash style, are common:
Typical gravity filtration flowrates range from 2 gpm/ft2–3 gpm/ft2 of media surface area. The minimum backwash flowrate is 30% of the process flow, but the flowrate requires adjustment with temperature changes. Anthracite is a light material that can be easily backwashed from the vessel as water density increases with colder temperatures.
Pressure filtration. Both vertical and horizontal pressure designs are available. Horizontal filters are sometimes used for large applications and are of cylindrical shape. The common diameter is 8 ft–10 ft, with lengths up to 40 ft. Many are equipped with a surface washer with a sweep arm.
This section focuses on vertical filters, similar to that shown in FIG. 7. A typical installation has multiple vessels to offer flexibility and redundancy. Filters can be as large as 12 ft in diameter, but per shipping constraints most filters are 10 ft or smaller in diameter. The design is a function of the required flowrate, media type, the number of filters and backwash configurations. For dual-media filters with only anthracite and sand, a common hydraulic guideline is 4 gpm/ft2–6 gpm/ft2: this range is common for potable water applications. A higher range of 5 gpm/ft2–8 gpm/ft2 is possible with multi-media filters containing anthracite, sand and garnet.
A well-designed multi-media filter might have 18 in. of anthracite (with a particle diameter range of 1 mm ± 0.2 mm) as the top layer with a void space of 50%. Next is 12 in. of sand (with a particle diameter range of 0.5 mm ± 0.1 mm) with a void space of 39%. Finally comes the heaviest media, 6 in. of garnet (with a particle diameter range of 0.25 mm ± 0.05 mm) with a void space of 47%. The media particle size decreases from top to bottom.
Note also the “intermixing zones” between media layers. These are usually 4 in.–6 in. in depth and should be expected, as some mixing of media always occurs. If a flow irregularity causes expansion or distortion of these zones, filtration efficiency could be affected.
Let us now examine flow calculations through the media. Assume that the filter in FIG. 7 has an influent flowrate of 5 gpm/ft2. The linear flowrate is determined from the following calculation (Eq. 9) using dimensional analysis:
5 min ÷ 7.48 = 0.67 ft/min (9)
Because the anthracite has a void space of 50%, the linear velocity doubles. The velocity increases to 3.44 ft/min when passing through sand, and finally steadies at 7.31 ft/min after passage through the garnet. This is a large velocity increase, and any delicate flocs that carry over from the clarifier may be broken into very fine solids that can potentially foul downstream equipment. Additional filtration may be needed to capture these residual fines and other particulates that could affect RO or ion exchange systems.
Filter inspection. Periodic filter inspections are necessary to ensure reliable performance. Three items are of particular importance:
Visual inspection. A visual inspection of the top surface can help to identify the amount of carryover to the media and if backwashing procedures are adequate. Visual observation, with measurement to the media surface, will reveal any loss of anthracite. Anthracite can be lost when the backwash flowrate is not adjusted for temperature. Also, coagulant or flocculant carryover from the clarifier can cause anthracite grains to agglomerate, increasing their buoyancy.
Depth sampling. Sampling devices are available that can be inserted from top to bottom of the media, and then extracted with a media column contained therein. Careful discharge of the sampler contents on a clean, flat surface will reveal media conditions (FIG. 8).
The large, vertical blue arrow points to the intermixing zone for the anthracite and sand and is within the recommended 4 in.–6 in. range. Similarly, the red arrow points to the intermixing zone for sand and garnet, and is also satisfactory. Depth sampling also allows visual evaluation of media cleanliness and condition.
Filter performance monitoring. Online turbidity monitoring (FIG. 9) is a common and effective method to analyze clarifier/filter performance.
A well-designed and operated clarifier/filter can produce effluent with a turbidity near 0.1 mS/cm. Periodic supplemental analyses can help to confirm performance. These include:
While the relationship of turbidity and TSS is variable depending on the water supply and treatment methods, periodic TSS sampling can provide a reasonable correlation for individual sites. Particle counts are not always included in a sampling program. Sampling for microorganisms helps to determine the effectiveness of biocide feed, and to determine the potential for equipment biofouling. This subject will appear again in Parts 2 and 3 of this series.
POTW treatment. Increasingly, either by mandate or choice, POTW effluent has become the raw water supply for industrial facilities. Often, this water has undergone only secondary treatment at the wastewater plant, which means that it still contains significant concentrations of ammonia and nitrite/nitrate, phosphate, organics, suspended solids and microbes. These constituents can initiate and exacerbate severe microbiological fouling in cooling systems and other water networks. Some of the compounds—most notably the nitrogen species and many of the organics—are not captured by conventional clarification/filtration. Supplemental microbiological treatment methods may be necessary to remove these impurities. Emerging are membrane bioreactors (MBR) and moving-bed bioreactors (MBBR). A basic schematic of the former is illustrated in FIG. 10.
The fundamental process is that of return activated sludge, in which beneficial microorganisms consume the organic compounds and nutrients that enter the main vessel from the mixing zone. A recycle stream helps to bring active, well-established organisms to the inlet of the mixing zone. A major difference of MBR from conventional activated sludge is the use of microfilter membranes rather than a traditional clarifier to separate solids from the effluent. The microfilters produce a very clear stream, essentially free of suspended solids—Micro- and ultrafiltration are addressed in Part 2 of this article (September 2024). One deficiency of the most basic MBR process is that ammonia is only converted to nitrite/nitrate, which can still serve as a nutrient for some micro-organisms. This issue can be corrected by expanding the MBR system to include anoxic or anaerobic reaction chambers containing microorganisms that convert nitrite/nitrate to elemental nitrogen.
MBBR employs an agitated vessel with mobile plastic media that circulates within the compartment. These media provide sites for the microbes to settle and develop healthy colonies. MBBR resembles a modern version of the old trickling bed technology for wastewater treatment. Unlike MBR, the micro- or ultrafilters that polish MBBR effluent must be located externally to the reaction vessel. HP
DISCLAIMER
The author has included much information in this article from consultation with Ed Sylvester, Director; Filtration, Ion Exchange & Membrane Technologies, ChemTreat Inc., Glen Allen, Virginia (U.S.), who focuses heavily on the refinery industry. However, the views represent those of the author. None of the programs or methods outlined in this series should be implemented without first contacting reputable manufacturers and consultants.
NOTES
LITERATURE CITED
Brad Buecker is President of Buecker & Associates LLC, and specializes in consulting and technical writing/marketing. Most recently, he served as a Senior Technical Publicist with ChemTreat Inc. Buecker has many years of experience in or supporting the power industry, much of it in steam generation chemistry, water treatment, air quality control, and results engineering positions with City Water, Light & Power (Springfield, Illinois, U.S.) and at the Kansas City Power & Light Company's (now Evergy) La Cygne, Kansas (U.S.) station. His work has also included 11 years with two engineering firms, Burns & McDonnell and Kiewit, and he spent two years as acting Water/wastewater Supervisor at a chemical plant. Buecker earned a BS degree in chemistry from Iowa State University with additional course work in fluid mechanics, energy and materials balances, and advanced inorganic chemistry. He has authored or co-authored > 250 articles for various technical trade magazines, and he has written three books on power plant chemistry and air pollution control. He is a member of the ACS, AIChE, AIST, ASME, AWT, CTI, the Electric Utility Chemistry Workshop planning committee, and is active with the International Water Conference and Power-Gen International.